A Unified Framework for Lifelong Learning in Deep Neural Networks

Humans can learn a variety of concepts and skills incrementally over the course of their lives while exhibiting an array of desirable properties, such as non-forgetting, concept rehearsal, forward transfer and backward transfer of knowledge, and so on. Previous approaches to lifelong learning (LLL) have demonstrated subsets of these properties, often with multiple mechanisms. In this paper, we propose a simple yet powerful unified framework that demonstrates all of these desirable properties. Our novel framework utilizes a small number of weight consolidation parameters dynamically applied to groups of weights, reflecting how "stiff" weights can be modified during training in deep neural networks. In addition, we are able to draw many parallels between the behaviours and mechanisms of our model and those surrounding human learning, such as memory loss or sleep deprivation. This allows our approach to serve as a conduit for two-way inspiration to further understand lifelong learning in machines and humans.

[1]  William D S Killgore,et al.  Effects of sleep deprivation on cognition. , 2010, Progress in brain research.

[2]  Shan Yu,et al.  Continual learning of context-dependent processing in neural networks , 2018, Nature Machine Intelligence.

[3]  Jason Weston,et al.  Curriculum learning , 2009, ICML '09.

[4]  D. Treffert,et al.  Inside the mind of a savant. , 2005, Scientific American.

[5]  Patrick Jähnichen,et al.  Learning to Remember: A Synaptic Plasticity Driven Framework for Continual Learning , 2019, 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[6]  Stefan Wermter,et al.  Continual Lifelong Learning with Neural Networks: A Review , 2019, Neural Networks.

[7]  Anthony V. Robins,et al.  Catastrophic Forgetting, Rehearsal and Pseudorehearsal , 1995, Connect. Sci..

[8]  Michael Carbin,et al.  The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks , 2018, ICLR.

[9]  Geoffrey E. Hinton,et al.  Distilling the Knowledge in a Neural Network , 2015, ArXiv.

[10]  Richard Socher,et al.  Learn to Grow: A Continual Structure Learning Framework for Overcoming Catastrophic Forgetting , 2019, ICML.

[11]  OctoMiao.  Overcoming catastrophic forgetting in neural networks , 2016 .

[12]  Christoph H. Lampert,et al.  iCaRL: Incremental Classifier and Representation Learning , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[13]  David Barber,et al.  Online Structured Laplace Approximations For Overcoming Catastrophic Forgetting , 2018, NeurIPS.

[14]  Susan M. Barnett,et al.  When and where do we apply what we learn? A taxonomy for far transfer. , 2002, Psychological bulletin.

[15]  Sung Ju Hwang,et al.  Lifelong Learning with Dynamically Expandable Networks , 2017, ICLR.

[16]  Sebastian Thrun,et al.  Lifelong Learning Algorithms , 1998, Learning to Learn.

[17]  Matthew P Walker,et al.  Sleep, memory and emotion. , 2010, Progress in brain research.

[18]  Zhanxing Zhu,et al.  Reinforced Continual Learning , 2018, NeurIPS.

[19]  Michael McCloskey,et al.  Catastrophic Interference in Connectionist Networks: The Sequential Learning Problem , 1989 .

[20]  Gerald Tesauro,et al.  Learning to Learn without Forgetting By Maximizing Transfer and Minimizing Interference , 2018, ICLR.

[21]  Marc'Aurelio Ranzato,et al.  Efficient Lifelong Learning with A-GEM , 2018, ICLR.

[22]  Philip H. S. Torr,et al.  Riemannian Walk for Incremental Learning: Understanding Forgetting and Intransigence , 2018, ECCV.

[23]  Gregory Cohen,et al.  EMNIST: an extension of MNIST to handwritten letters , 2017, CVPR 2017.

[24]  Charles X. Ling,et al.  Region-Based Global Reasoning Networks , 2020, AAAI.

[25]  Surya Ganguli,et al.  Continual Learning Through Synaptic Intelligence , 2017, ICML.

[26]  Marc'Aurelio Ranzato,et al.  Gradient Episodic Memory for Continual Learning , 2017, NIPS.

[27]  P. S. St George-Hyslop,et al.  Prediction of probable Alzheimer's disease in memory-impaired patients , 1996, Neurology.

[28]  Marcus Rohrbach,et al.  Memory Aware Synapses: Learning what (not) to forget , 2017, ECCV.

[29]  Charles X. Ling,et al.  Pelee: A Real-Time Object Detection System on Mobile Devices , 2018, NeurIPS.

[30]  David Filliat,et al.  Don't forget, there is more than forgetting: new metrics for Continual Learning , 2018, ArXiv.